MARINE ECOLOGY PROGRESS SERIES Vol. 60: 299303. 1990 Published February 22 Mar. Ecol. Prog. Ser.

NOTE

Coral mortality associated with blooms in the eastern Pacific (Costa Rica and Panama)

Hector M. ~uzman',Jorge cortes2,*,Peter W. ~lynn~,Robert H. ~ichmond~

' Smithsonian Tropical Research Institute (Panama). APO Miami, Florida 34002-001 1, USA * Division oi Biology and Living Resources, Rosenstiel School of Marine and Almospheric Science, Universily of Miami, 4600 Rickenbacker Causeway, Miami, Florida 33149-1098, USA Marine Laboratory. University of Cuam, UOG Station, Mangilao, Guam 96923, USA

ABSTRACT: Coral reefs at Cano Island, Costa hca, and Uva tions, oxygen depletion, and hydrogen sulfide poison- Island. Panama, were affected during severe dinoflagellate ing rather than by toxins of Ptychodiscus brevis (Davis) blooms in 1985. In the second half of 1985. mass mortality of that were present during the event. reef fishes and invertebrates, especially reef corals, occurred during blooms of the Cochlodinium catenatum In this note we present observations of coral death and Gonyaulax monilata. At Cano Island, up to 100% coral associated with dinoflagellate blooms in the eastern mortality was observed between the surface and 3 m depth, Pacific - at Cano Island, Costa Rica and Uva Island, with pocilloporid species and Tubastrea coccinea most Panama - and discuss possible mechanisms that may severely affected. At Uva Island, only 13 % pocilloporid mor- tality occurred and this was confined to the shallowest reef have resulted in coral death. areas ('-3 m). The copious amounts of mucus associated with Study site and methods. Cano Island (8"43'N, C. catenatum, present in the water column to 1-3 m depth, 83"52'W) is situated in the southern region of the and adhering to coral colonies, suggested that coral death was Pacific coast of Costa hca, ca 15 km offshore (Fig. 1). caused by smothering. Other conditions that may have Information on the coral community structure and affected reef organisms were the presence of toxlc G. monilata, and possible oxygen depletion due to the hlgh climatic and oceanographic conditions is available in densities of phytoplankton and the decomposition of dead Guzman & Cortes (1989a, b). The percentage of dead organisms. These dinoflagellate blooms have interfered with colonies of all coral species encountered at Cano Island recovery of reefs disturbed dur~ngthe 1982/83 El Nino warm- in July 1985 was recorded. Eight permanent plots of ing event. 10 m2 each at 4 localities from 0.5 to 12 m depth that have been surveyed since January 1984 (for an assess- There are only 2 reports of coral death associated ment of coral recovery after the 1982/83 El Niiio event) with phytoplankton blooms. Baas-Becking (1951) were mapped in January 1985 before the dinoflagellate reported on the death of corals in the shallow waters of bloom and in late August 1985 near the end of the , and attributed this to the putrefaction bloom. Water around the reefs was collected using 11 of thick masses of causing a bottles in July 1985 and the species composition and marked reduction in dissolved oxygen. Smith (1975) relative abundances of the phytoplankters were deter- commented on the impact of a 1971 'red tide' on reef mined using a Palmer-Maloney chamber (Steidinger coral communities off the west coast of Florida, USA, 1979). with near-complete mortality of shallow-water (< 40 m) Uva Island reef (7"49'N, 81°46' W) is located about biotas over extensive areas. Smith hypothesized these 40 km offshore in the Gulf of Chiriqui, Panama, with mortalities were caused by bacterial and fungal infec- similar climatic conditions to those at Cano Island. Detailed information on the oceanographic conditions ' Permanent address: Centro de Investigacion en Ciencias and coral community structure are available in Dana del Mar y Limnologia (CIMAR),Universidad de Costa Rica, (1975) and Glynn (1976). The same scleractinian corals San Pedro, Costa Rica occur in both study areas; however, at Uva Island the

O Inter-Research/Printed in F. R. Germany 22 2 2 S'n g2 S g d.

$2 2 3 i? L $9 2 : 0 %v, 0- 23 2 m 3 v, h 2. cm 8. S tl D 'P m % 3 5 m, 2 $ V,Y

m %m zm n z m z "z " g gS? S?- EC mP 7 j. CY

+A P P P 0-W N N- WL~L-WCn N~C~OWNNC~- I I II l 3 F +m m+ 'P -2 0 d3LL.L 2odsLdsi.L L m 0 Cn Er

m z X P +NW + - +-4ww-s $2 4-UI cn U3W W- 4WwW wCnmr4mUIUI 5'5 g 2. 2

)-L ru Cn W 0 W wwwo 00UI 0 0000 00WO oowr+c.o,o g8 P G117n?anet al.. Coral mortal~ty 301

relative abundance of pocilloporid corals is higher than (0.5 to 3 m depth), while the massive corals remained at Cano Island. Ten permanent chain transects, unchanged (Table 2). The deeper stations (6 to 12 m between 3 and 6 m depth, were sampled at Uva Island depth) showed no significant change in live coral cover 3 mo before and 2 ino after the dinoflagellate bloom. between January and August 1985 (Table 2). The chain ti-ansects were 10 m long with 73 sampling The Uva Island dinoflagellate bloom was observed in points (links) m-' and the distance covered by each October and November 1985. It had a red-brown color species was recorded (Glynn 1984).The dinoflagellates and probably lasted several days Numerous pocillo- at Uva Island were not identified to species. porid corals at 1 to 2 m depth had bleached and were Results. A severe dinoflagellate bloom, or succession sloughing tissues. The CO-occurrence of viscous foam of blooms, occurred around Cano Island from 3 June to and dinoflagellates suggested that the former was pro- 12 July 1985, with additional pulses through October duced by the latter. The viscous foam near the surface (park rangers daily report, Cano Island Biological oscillated vertically with the tides over shallow-water Reserve). Surface waters, from 2 to 3 m depth, were corals, which were most affected. Transect sampling red-yellow with viscous foam presumably produced by 3 mo before the dinoflagellate bloom and 2 mo after the dinoflagellates. A surface water sample collected at indicated minor coral mortality in the shallowest trans- Cano Island in July 1985 contained 8.3 X 10' live ects (3 m). In 4 transects at 3 m depth, cover of live

phytoplankton cells 1-' and more than 3 X 106 cells I-' Pocillopora spp. was reduced by 12.8 O/O (Table 3), but including dead cells. The relative abundances by cell the decline was not statistically significant. The number were: Cochlodinium catenatum Okamura increases in coral cover in the deeper transects (97.0 %), Gonyaulax monilata Howell (1.4 %), (Table 3, Transects 5 to 8) were due to fragmentation (1.2 %) and other dinoflagellates (0.4 %). Numerous and the spread of live branches into the sampling area. fish were affected, with hundreds of scarids, balistids, Reductions also resulted from coral fragmentation and acanthurids, pomacentrids and tetraodontids found the transport of pocilloporid corals away from a transect dead on the beach, as well as dead hermit crabs, (Table 3, Transect 9) and the burial of massive corals brachyuran crabs and gastropods. Some shallow water by dead pocilloporid branches (Table 3, Transect 10). corals were also affected (Table l), especially Pocillo- Discussion. The 1985 dinoflagellate bloom was the pora elegans Dana, P. damicornis (Linnaeus) and the most severe reported off the Pacific coast of Costa Rica azooxanthellate coral Tubastrea coccinea Lesson. The during the last 15 yr (R. Viquez pers. comm. 1988). presence of live obligate crustacean symbiotes in some Intense upwellings were experienced that year in the dead pocilloporid corals indicated recent mortality. A Gulf of Panama, which were related to anti-El Nino significant inverse relation was found between the type activity (Glynn 1989). These conditions could have number of dead colonies and depth (tau = -0.923, p caused a general shoaling of the nutricline in non- < 0 03, n = 20, Kendall rank correlation). No recently upwelling areas, ultimately leading to the dinoflagel- dead or partially dead corals were observed in deeper late bloom observed in 1985. waters (2 10 m). In previous years at the Gulf of Nicoya, less than Between January and August 1985, the percentage 100 km northwest of Cano Island, blooms were also cover of live pocillopond corals in the permanent plots dominated by Cochlodinlum catenatum, according to at Cano Island decreased to zero at the shallow stations Hargraves & Viquez (1981, 1985). This species has not

Table 2 Percent live hermatypic coral cover hefore and after a d~noflagellatebloom In 10 m2 plots, Cano Island, Costa Rica. Pocilloporid corals were Pocillopora elegans and Poallopora damicornis; other corals were Pavona varians, Pavona gigantea, Pavona clavus, Psammocora stellata, Gardlneroseris planulata and Porites lobata

Plot Depth 19 January 1985 29 August 1985 no (m) Pocillopond corals All corals Pocilloporid corals All corals

1 0.5-1 24.19 27.00 0 3.35 2 0.5-1 30.19 30.81 0 173 3 2-3 34.60 37.50 0 2.61 4 2-3 36.90 40 40 0 3.39 5 6 0 30.40 0 27 47 6 6 0 31.57 0 27.50 7 10-12 5.13 25.48 4.25" 23 35 8 10-12 5.31 20.16 4.91a 16 54

" Reduction due to predation by Acanthaster plancl 302 Mar. Ecol. Prog. Ser. 60: 299-303. 1990

Table 3. Percent live hermatypic coral cover before and after a phytoplankton bloom, Uva Island, Gulf of Chiriqui, Panama. Pocilloporid corals were Pocillopora darnicornis and P. elegans; other corals were Gardinerosens planulata, Pavona varians and Psammocora stellata

Transect Depth 13 August 1985 18 January 1986 no. (m) Pocilloporid corals All corals Pocilloporid corals All corals

1 3 25.9 25.9 20.7 20.7 2 3 21.5 21.5 21.0 21.0 3 3 23.6 23.6 20.6 20.6 4 3 22.5 22.9 18.9 18.9 5 4 36.8 37.1 40.8 42.0 6 4 19.3 19.7 23.8 24.1 7 5 14.7 14.7 20.8 20.8 8 5 23.7 25.8 24.9 25.5 9 5 14.4 14.4 8.5 8.5 10 6 0 27.1 0 19.7

been proven to be toxic, but other species of Coch- Betzer 1975), oxygen stress may occur on reefs, espe- lodinium were toxic to fish in Japan (Yuki & Yoshi- cially in the presence of decomposing organisms. matsu 1989). Even though concentrations of dinoflag- An increase in viscosity of the water due to mucus ellates during those blooms were higher than at Cano produced by dinoflagellates has been suggested as a Island, Hargraves & Viquez (1981, 1985) did not report possible cause of death of fish (Jenkinson 1989). There any massive mortality in the estuarine Gulf of Nicoya. are no reports of coral death attributed to smothering This difference in mortality between Nicoya and Cano by thick mucous layers due to phytoplankton blooms. may have been due to a greater susceptibility of corals However, a thick layer of mucus could effectively iso- to changes in water quality than most estuarine organ- late corals from their environment, preventing feeding, isms and because Gonyaulaxmonilata was not found at respiration and excretion, and if the blockage were Nicoya. prolonged, the coral would die. Also, some The possible mechanisms by which phytoplankton thrive in the mucus produced by corals and an increase blooms cause mortality include: (a) toxicity, (b) reduc- in bacterial density could cause coral death by sulfide tion of light penetration, (c) depletion of dissolved oxy- poisoning, oxygen depletion and attack of coral tissue gen, (d) smothering by mucus, or (e) any combination by the bacteria (Ducklow & Mitchell 1979, Paul et al. of the above effects. 1986). Mortality caused by toxic dinoflagellate blooms is Eastern Pac~ficcoral reefs are recovenng slowly from well documented from boreal to tropical seas the 1982/83 El Nifio warming event that caused mass (Steidinger & Baden 1984, Anderson et al. 1985). Of the coral mortalities in Panama and Costa Rica (Glynn et al. dinoflagellates observed in the blooms, only Gony- 1988). The 1985 dinoflagellate blooms occurred less aulax monilata is known to be toxic (Steidinger & than 2 yr after this disturbance. At Uva Island, where Baden 1984), but even at low densities its toxicity may the El Nino disturbance resulted in 70 to 95'~coral cause the death of organisms (Schantz et al. 1975 in mortality, few corals were affected by the dinoflagel- Steidinger 1983). late bloom except at shallow depths (5 3 m). At Cano Light reduction during the dirioflagellate blooms did Island, El Nilio-associ.ated coral mortality was lower, 25 not seem to be critical in the death of corals. Two to 75 %, but mortality due to the dinoflagellate bloom observations suggest this: (1) the azooxanthellate coral was high, completely eliminating some coral species Tubastrea coccinea, which is not light dependent, was from shallow reef zones. significantly affected; and (2) corals in deeper water In conclusion, the death of reef organisms at Cano were not affected. and Uva Islands was possibly caused by a combination Death of marine organisms attributed to anoxia after of (a) toxicity, (b) oxygen depletion, and (c) smothering phytoplankton blooms has been reported in several by mucus produced during dinoflagellate blooms. marine habitats, including coral reefs (Baas-Becking Adhesion of mucus to polypal areas, and the interfer- 1951, Smith 1975) The majority of the fishes and crabs ence with polyp expansi.on, seemed the most likely affected at Caiio and Uva Islands were active, free- cause of mortality in corals. These dinoflagellate living members of the epibenthos. Also, as oxygen blooms, occurring 2 yr after the 1982/83 El Nino tvarm- levels in warm waters are often close to the lower ing event, have slowed the recovery of these disturbed tolerance limits of aerobic organisms (Johannes & reefs. Guzman et al.: Coral mortality 303

Acknowledgen~elzts.Research supported by NSF grant OCE Guzman, H. M,, Cortes, J (1989b). Growth rates of eight 8415615 to P.W G. We thank R. Viquez, Universidad Nacional, species of scleractinlan corals in the eastern Pac~fic(Costa Heredia, Costa Rica, for the identification of dinoflagellates Rica) Bull. mar. Sci 44: 1186-1194 and L. Brand, J. Cubit, C. M. Eakin, P. Hallock, R. Viquez, Hargraves. P. E., Viquez. R. (1981). The dinoflagellate red G. M. Wellington and 3 anonymous reviewers for criticizing tide in Golfo de Nicoya. Costa Rica. Rev. Biol. Trop. 29: the manuscript. 31-38 Hargraves. P. E., \liquez. R. (1985). Spat~aland temporal LITERATURE CITED distr~but~onof phytoplankton in the Gulf of Nicoya, Costa Rlca Bull. mar. Sci. 37: 577-585 Anderson, D M, White, A. W., Baden, D. G. (eds) (1985). Jenklnson, I R. (1989).Increases in viscosity may kill fish in Toxic dinoflagellates. Proceedings of the 3rd internatlonal some blooms. In: Okaichl, T., Anderson, D M.,Nemoto, T conference on toxic dinoflagellates, St. Andrews, New (eds.) Red tides: biology, env~ronmentalscience, and tox- Brunswick, Canada. Elsevier, New York icology. Elsevier, New York, p. 435-438 Baas-Becking. L. G. M. (1951).Notes on some Cyanophyceae of Johannes, R. E., Betzer. S. B. (1975). Introduction: marine the Pacific Region. Proc. K. Ned. Akad. Wet. Ser. C 54: communities respond differently to pollution in the tropics 213-225 than at hlgher latitudes. In- Ferguson-Wood, E. J., Johan- Dana, T F (1975). Development of contemporary eastern nes, R E (eds.) Tropical marlne pollution. Elsevier, New Pacific coral reefs. Mar Biol 33 355-374 York, p 1-12 Ducklow. H. W., Mitchell. R. (1979) Bacterial populations and Paul, J. H., DeFlaun, M. F, Jeffrey, W H. (1986). Elevated adaptations in the mucus layers on living corals. Limnol. levels of microbial activ~tyin the coral surface microlayer Oceanogr 24: 715-725 Mar Ecol. Prog. Ser 33: 2940 Glynn, P. W (1976). Some physical and biological determi- Smith. G. B. (1975).The 1971 red tide and its impact on certain nants of coral community structure in the eastern Pacific. reef communities in the mid-eastern Gulf of Mexico. Envi- Ecol. Monogr. 46: 431-456 ron Lett 9: 141-152 Glynn, P. W (1984).Widespread coral mortality and the 1982/ Steidlnger, K. A. (1979).Collection, enumeration and identifi- 83 El N~riowarming event Env~ron.Conserv. 11: 133-146 cation of free-living marine dinoflagellates In: Taylor, D. Glynn, P. \V (1989). Coral mortality and disturbances to reef L., Seliger, H. H. (eds.) Toxic dinoflagellate blooms. corals in the tropical eastern Pacific. In: Glynn, P. W. (ed.) Elsevier/North-Holland, New York, p. 435442 Global ecological consequences of the 1982-1983 El Nino- Steidinger. K. A. (1983).A re-evaluation of toxic dinoflagellate Southern Oscillation. Elsevier, Amsterdam, p. 55-126 biology and ecology. Prog. phycol. Res. 2: 147-188 Glynn, P. W, Cortes, J., Guzman. H. M., Richmond. R. H. Steidinger. K. A., Baden, D. G. (1984). Toxic marine dino- (1988) El Nino (1982-83) associated coral mortality and flagellates. In. Spector, D L. (ed.) Dinoflagellates. relationship to sea surface temperature deviat~onsin the Academ~cPress, Orlando, p 201-261 tropical eastern Pacific. Proc 6th lnt. Coral Reef Symp., Yuki, K., Yoshimatsu, S. (1989). Two fish-killlng species of Townsv~lle,in press Cochlodinium from Harlma Nada, Seto Inland Sea, Japan. Guzman, H. bf., Cortes, J. (1989a). Coral reef community In: Okaichi, T., Anderson. D. M,, Nemoto, T (eds.) Red structure at (:clno Island, Pacific Costa Rica. P.S.Z.N.I. Mar tides: biology, environmental science, and toxicology. Ecol. 10: 2143 Elsevier. New York, p. 451454

This note was presented by Professor R. P. M. Bak, Texel, Manuscript first received: July 19, 1989 The Netherlands Revlsed version accepted: Noveniber 21, 1989